A short DICOM toy example for how to create a missing DICOM sequence.
We have one DICOM files that 'works', e.g. can be used for processing in a third-party application. Now we have a new scanner (combined MR/PET) that stores PET related information in different fields in the DICOM meta-data. Here is an example of what the sequence information looks like that our software depends on (created with dcmdump +P 0054,0016 ok_file.dcm)
(0054,0016) SQ (Sequence with explicit length #=1) # 470, 1 RadiopharmaceuticalInformationSequence
(fffe,e000) na (Item with explicit length #=17) # 462, 1 Item
(0009,0010) LO [MEDISO-1] # 8, 1 PrivateCreator
(0009,10ee) DT [20200305100500.685] # 18, 1 Unknown Tag & Data
(0009,10ef) DT [20200305111300.303] # 18, 1 Unknown Tag & Data
(0009,10f0) FD 19.5 # 8, 1 Unknown Tag & Data
(0009,10f1) FD 0.9 # 8, 1 Unknown Tag & Data
(0009,10f2) FD 0.2 # 8, 1 Unknown Tag & Data
(0009,10f3) FD 0 # 8, 1 Unknown Tag & Data
(0009,10fa) ST [MBq] # 4, 1 Unknown Tag & Data
(0009,10fb) US 1 # 2, 1 Unknown Tag & Data
(0018,1071) DS [0.2] # 4, 1 RadiopharmaceuticalVolume
(0018,1074) DS [8100000.500000] # 14, 1 RadionuclideTotalDose
(0018,1075) DS [6586.2] # 6, 1 RadionuclideHalfLife
(0018,1076) DS [0.9673] # 6, 1 RadionuclidePositronFraction
(0018,1078) DT [20200305100928.053] # 18, 1 RadiopharmaceuticalStartDateTime
(0054,0300) SQ (Sequence with explicit length #=1) # 56, 1 RadionuclideCodeSequence
(fffe,e000) na (Item with explicit length #=3) # 48, 1 Item
(0008,0100) SH [C-111A1] # 8, 1 CodeValue
(0008,0102) SH [SRT] # 4, 1 CodingSchemeDesignator
(0008,0104) LO [^18^Fluorine] # 12, 1 CodeMeaning
(fffe,e00d) na (ItemDelimitationItem for re-encoding) # 0, 0 ItemDelimitationItem
(fffe,e0dd) na (SequenceDelimitationItem for re-encod.) # 0, 0 SequenceDelimitationItem
(0054,0302) SQ (Sequence with explicit length #=1) # 60, 1 AdministrationRouteCodeSequence
(fffe,e000) na (Item with explicit length #=3) # 52, 1 Item
(0008,0100) SH [G-D101] # 6, 1 CodeValue
(0008,0102) SH [SRT] # 4, 1 CodingSchemeDesignator
(0008,0104) LO [Intravenous route] # 18, 1 CodeMeaning
(fffe,e00d) na (ItemDelimitationItem for re-encoding) # 0, 0 ItemDelimitationItem
(fffe,e0dd) na (SequenceDelimitationItem for re-encod.) # 0, 0 SequenceDelimitationItem
(0054,0304) SQ (Sequence with explicit length #=1) # 68, 1 RadiopharmaceuticalCodeSequence
(fffe,e000) na (Item with explicit length #=3) # 60, 1 Item
(0008,0100) SH [C-B1031] # 8, 1 CodeValue
(0008,0102) SH [SRT] # 4, 1 CodingSchemeDesignator
(0008,0104) LO [Fluorodeoxyglucose F^18^] # 24, 1 CodeMeaning
(fffe,e00d) na (ItemDelimitationItem for re-encoding) # 0, 0 ItemDelimitationItem
(fffe,e0dd) na (SequenceDelimitationItem for re-encod.) # 0, 0 SequenceDelimitationItem
(fffe,e00d) na (ItemDelimitationItem for re-encoding) # 0, 0 ItemDelimitationItem
(fffe,e0dd) na (SequenceDelimitationItem for re-encod.) # 0, 0 SequenceDelimitationItem
Here the same information using dcm2xml:
<sequence tag="0054,0016" vr="SQ" card="1" len="470" name="RadiopharmaceuticalInformationSequence">
<item card="17" len="462">
<element tag="0009,0010" vr="LO" vm="1" len="8" name="PrivateCreator">MEDISO-1</element>
<element tag="0009,10ee" vr="DT" vm="1" len="18" name="Unknown Tag & Data">20200305100500.685</element>
<element tag="0009,10ef" vr="DT" vm="1" len="18" name="Unknown Tag & Data">20200305111300.303</element>
<element tag="0009,10f0" vr="FD" vm="1" len="8" name="Unknown Tag & Data">19.5</element>
<element tag="0009,10f1" vr="FD" vm="1" len="8" name="Unknown Tag & Data">0.9</element>
<element tag="0009,10f2" vr="FD" vm="1" len="8" name="Unknown Tag & Data">0.2</element>
<element tag="0009,10f3" vr="FD" vm="1" len="8" name="Unknown Tag & Data">0</element>
<element tag="0009,10fa" vr="ST" vm="1" len="4" name="Unknown Tag & Data">MBq</element>
<element tag="0009,10fb" vr="US" vm="1" len="2" name="Unknown Tag & Data">1</element>
<element tag="0018,1071" vr="DS" vm="1" len="4" name="RadiopharmaceuticalVolume">0.2</element>
<element tag="0018,1074" vr="DS" vm="1" len="14" name="RadionuclideTotalDose">8100000.500000</element>
<element tag="0018,1075" vr="DS" vm="1" len="6" name="RadionuclideHalfLife">6586.2</element>
<element tag="0018,1076" vr="DS" vm="1" len="6" name="RadionuclidePositronFraction">0.9673</element>
<element tag="0018,1078" vr="DT" vm="1" len="18" name="RadiopharmaceuticalStartDateTime">20200305100928.053</element>
<sequence tag="0054,0300" vr="SQ" card="1" len="56" name="RadionuclideCodeSequence">
<item card="3" len="48">
<element tag="0008,0100" vr="SH" vm="1" len="8" name="CodeValue">C-111A1</element>
<element tag="0008,0102" vr="SH" vm="1" len="4" name="CodingSchemeDesignator">SRT</element>
<element tag="0008,0104" vr="LO" vm="1" len="12" name="CodeMeaning">^18^Fluorine</element>
</item>
</sequence>
In order to read the dose information from the ok_file we can use dcmdump:
# read dose information from input
#(0018,1074) DS [8100000.500000] # 14, 1 RadionuclideTotalDose
#(0018,1075) DS [6586.199707] # 12, 1 RadionuclideHalfLife
#(0018,1076) DS [0.970000] # 8, 1 RadionuclidePositronFraction
#(0018,1078) DT [20211206134309] # 14, 1 RadiopharmaceuticalStartDateTime
#(0018,0031) LO [F-18] # 4, 1 Radiopharmaceutical
radiopharma=$( dcmdump +P 0018,0031 "$input" | cut -d'[' -f2 | cut -d']' -f1)
echo "Radiopharmaceutical (from input file $input): $radiopharma"
dosetotal=$( dcmdump +P 0018,1074 "$input" | cut -d'[' -f2 | cut -d']' -f1)
echo "Totaldose (from input file $input): $dosetotal"
halflife=$( dcmdump +P 0018,1075 "$input" | cut -d'[' -f2 | cut -d']' -f1)
echo "Halflife (from input file $input): $halflife"
positronfraction=$( dcmdump +P 0018,1076 "$input" | cut -d'[' -f2 | cut -d']' -f1)
echo "Positron fraction (from input file $input): $positronfraction"
startdate=$( dcmdump +P 0018,1078 "$input" | cut -d'[' -f2 | cut -d']' -f1)
echo "Start date (from input file $input): $startdate"
Adding this information into a sequence item we need to use the path notation from dcmodify.
# add this value to the template.dcm file from which we will copy the values
dcmodify -m "(0054,0016)[0].(0018,1074)=$dosetotal" template.dcm
# we would want to run this - but this does not work (copy a whole sequence):
# dcmodify -if "(0054,0016)[0].(0018,0015)=template.dcm" test2.dcm
# instead we do this item by item
dd="dicom.dic"
# add sequence for 0054,0016 RadiopharmaceuticalInformationSequence
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,0010)=MEDISO-1" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10ee)=20200305100500.685" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10ef)=20200305111300.303" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10f0)=19.5" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10f1)=0.9" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10f2)=0.2" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10f3)=0" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10fa)=MBq" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0009,10fb)=1" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0018,1071)=0.2" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0018,1074)=8100000.500000" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0018,1075)=6586.2" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0018,1076)=0.9673" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0018,1078)=20200305100928.053" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0054,0300)[0].(0008,0100)=C-111A1" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0054,0300)[0].(0008,0102)=SRT" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0054,0300)[0].(0008,0104)=^18^Fluorine" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0054,0302)[0].(0008,0100)=G-D101" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0054,0302)[0].(0008,0102)=SRT" "$input"
DCMDICTPATH=$dd dcmodify -nb -i "(0054,0016)[0].(0054,0302)[0].(0008,0104)=Intravenous route" "$input"
There is one remaining issue. The tags in the 0009 group are private tags and therefore not in our default dcmtk data dictionary (dicom.dic). In order to get the correct value representation for all tags in the sequence these group-, tag pairs have been added to the standard dicom.dic as:
(0009,"MEDISO-1",10F0) FD PrivateCreator 1 PrivateTag
(0009,"MEDISO-1",10FA) ST SomethingWITHUnits 1 PrivateTag
(0009,"MEDISO-1",10F1) FD PrivateCreator 1 PrivateTag
(0009,"MEDISO-1",10F2) FD PrivateCreator 1 PrivateTag
(0009,"MEDISO-1",10F3) FD PrivateCreator 1 PrivateTag
(0009,"MEDISO-1",10EE) DT PrivateCreator 1 PrivateTag
(0009,"MEDISO-1",10EF) DT PrivateCreator 1 PrivateTag
(0009,"MEDISO-1",10FB) US PrivateCreator 1 PrivateTag
(0018,"MEDISO-1",1071) DS PrivateCreator 1 PrivateTag
(0018,"MEDISO-1",1074) DS PrivateCreator 1 PrivateTag
(0018,"MEDISO-1",1075) DS PrivateCreator 1 PrivateTag
(0018,"MEDISO-1",1076) DS PrivateCreator 1 PrivateTag
(0018,"MEDISO-1",1078) DS PrivateCreator 1 PrivateTag
Notice that each field is separated by a tab-character from the next.
The above steps of creating a sequence are available as a shell script cpyHeader.sh that also checks its input arguments (input folder and output folder) and runs on all found DICOM files.
TODO: The cpyHeader.sh script does not do the whole job currently. It is reading the dose information but not using that information when it writes the sequence. The resulting DICOM files have also not been tried yet with the analysis software. The individual dcmodify calls can be combined in fewer calls by repeatedly using the '-i' option. The limit there is just the maximum length of a command line, which includes the path to the DICOM files.